Re: table corrupted

From: "Joshua D(dot) Drake" <jd(at)commandprompt(dot)com>
To: João Eugenio Marynowski <joaoem(at)gmail(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: table corrupted
Date: 2009-10-22 23:16:07
Message-ID: 1256253367.2821.10.camel@jd-desktop.unknown.charter.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thu, 2009-10-22 at 14:28 -0200, João Eugenio Marynowski wrote:
> Hi
>
Repair? Not likely. Get past? Maybe.

set zero_damaged_pages to on;
vacuum verbose;

I would do a hardware check too.

Joshua D. Drake

> Can someone help me how to repair the problem below, I'm using
> Postgres 8.2.5:
> - after appeared the erros below in selects, vacuum and dump in one
> table:
> 2009-10-16 16:07:06 BRT 192.168.0.87 ERROR: could not access status
> of transaction 29024764
> 2009-10-16 16:07:06 BRT 192.168.0.87 DETAIL: Could not open file
> "pg_clog/001B": No such file or directory.
> 2009-10-16 16:07:06 BRT 192.168.0.87 STATEMENT: select ...
> 2009-10-16 16:11:47 BRT 192.168.0.29 ERROR: invalid page header in
> block 462821 of relation "..."
> 2009-10-16 16:11:47 BRT 192.168.0.29 STATEMENT: select ....
> I created the file pg_clog/001B with 256kB of /dev/zero
> That resolve the problem with vacuum but began other error in selects
> and dump to the same table ended all connections and stay up after
> showing the error:
> 2009-10-19 13:50:03 BRT LOG: server process (PID 1544) was
> terminated by signal 11
> 2009-10-19 13:50:03 BRT LOG: terminating any other active server
> processes
> 2009-10-19 13:50:03 BRT 192.168.0.253 WARNING: terminating connection
> because of crash of another server process
> 2009-10-19 13:50:03 BRT 192.168.0.253 DETAIL: The postmaster has
> commanded this server process to roll back the current transaction and
> exit, because another server process exited abnormally and possibly
> corrupted shared memory.
> 2009-10-19 13:50:03 BRT 192.168.0.253 HINT: In a moment you should be
> able to reconnect to the database and repeat your command.
> Was habilited the zero_damage_pages option then executed selects,
> vacuums, and dumps but not changed...
> Was identified 2 register that if refered cause error.
> The BD was restored in backup server with 8.2.7 and executed vacuums
> ok but select and reindex crashed...
>
> Instaled 8.3.8 version and used pg_dump but error
> And then the select below show the problem where the codentrega from
> where clause differ from select answer:
>
> LOGIST=# select codentrega from entregas where codentrega='9879622';
> codentrega
> ------------
> z879622
> (1 registro)
>
> Any idea?

--
PostgreSQL.org Major Contributor
Command Prompt, Inc: http://www.commandprompt.com/ - 503.667.4564
Consulting, Training, Support, Custom Development, Engineering
If the world pushes look it in the eye and GRR. Then push back harder. - Salamander

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Magnus Hagander 2009-10-22 23:59:19 Re: client_lc_messages
Previous Message Tom Lane 2009-10-22 23:12:58 Re: plpgsql EXECUTE will not set FOUND